AI unemployment is still at zero.

But 22-year-olds are stopping to find work.

And this is the signal nobody is reading.

Not from some random TikToker.

From a study published directly by Anthropic, the company that built Claude.

📌 Translation: the people building these models already know what's happening to the labor market. And they decided to measure it, publish it, and sign it.

This is the data point that stopped me:

Among workers aged 22-25, in the professions most exposed to AI, hiring has dropped 14% since 2022.

14%.

Not layoffs.

Hires that never happen.

And the unemployment rate?

Flat. Unchanged. Zero signal.

Why?

Because people who don't get hired don't show up in unemployment statistics.

They disappear. They stay out. They don't get counted.

📌 This is exactly the kind of shift I talked about in the latest issue of AI Espresso: the economic model that justified paying you can collapse while the numbers still look normal.

And the most at-risk professions?

Not the ones you'd expect.

↳ Programmers

↳ Customer service reps

↳ Financial analysts

↳ Data entry

Educated profiles. Well paid. Often with a master's degree.

The cook, the mechanic, the bartender? Near-zero AI coverage.

Who's most at risk isn't who works with their hands.

It's who works with their head.

And here's the paradox that stings:

AI hasn't "stolen" any job in a measurable way yet.

But it's already changing who gets hired.

And young people entering the market right now, in certain sectors, are finding closed doors they don't even recognize as closed.

This isn't science fiction.

It's already in the data. Published by the people building the models.

📬 In the latest issue of AI Espresso I went deeper on this: why your job can evaporate while the statistics still look fine, and what to build so you don't find yourself outside the door without even noticing.